Table of Contents
- 1 Introduction
- 2 Imports
- 3 Load the final text cleancat15 data
- 4 Plot g-squared for monochromatic and chromatic files
- 5 Contour Plots
- 6 Now, Take only the data above the boundary as cleaned data
- 7 gm vs gc Plots
- 8 Remove outliers based on gm vs gc plot
- 9 Plot magnidude for different bin numbers
- 10 Magnitude weight column for Monochromatic case
- 11 Magnitude weight column for Chromatic case
- 12 Ellipticity Components Transformation
- 13 Radial Bins groupings
- 14 Plot of eT for mono and chro
- 15 Interactive plots
- 16 ipywidgets
- 17 Time Taken
Kernel Author:
Bhishan Poudel, Ph.D Contd. Astrophysics
Update: Jan 17, 2020 Fri
Date : Jan 13, 2020
Bhishan Poudel, Ph.D Contd. Astrophysics
Update: Jan 17, 2020 Fri
Date : Jan 13, 2020
Introduction¶
Update
- Looked at gm0 vs gc0 (and gm1 vs gc1) 45 degree line and removed outliers.
- Find the weights for g_sq for given magnitude bins using smooth fitting curve.
Usual Filtering
df = df.query('calib_psfCandidate == 0.0')
df = df.query('deblend_nChild == 0.0')
df['ellip'] = np.hypot( df['ext_shapeHSM_HsmShapeRegauss_e1'] ,
df['ext_shapeHSM_HsmShapeRegauss_e2'] )
df = df.query('ellip < 2.0') # it was 1.5 before
#select only few columns after filtering:
cols_select = ['base_SdssCentroid_x', 'base_SdssCentroid_y',
'base_SdssCentroid_xSigma','base_SdssCentroid_ySigma',
'ext_shapeHSM_HsmShapeRegauss_e1','ext_shapeHSM_HsmShapeRegauss_e2',
'base_SdssShape_flux']
df = df[cols_select]
# drop all nans
df = df.dropna()
# additional columns
df['radius'] = df.eval(""" ( (ext_shapeHSM_HsmSourceMoments_xx * ext_shapeHSM_HsmSourceMoments_yy) \
- (ext_shapeHSM_HsmSourceMoments_xy**2 ) )**0.25 """)
Shape filtering
https://github.com/LSSTDESC/DC2-analysis/blob/master/tutorials/object_gcr_2_lensing_cuts.ipynb
df = df.query('ext_shapeHSM_HsmShapeRegauss_resolution >= 0.3')
df = df.query('ext_shapeHSM_HsmShapeRegauss_sigma <= 0.4')
df = df.query('ext_shapeHSM_HsmShapeRegauss_flag== 0.0')
Filter strongly lensed objects
- Take the objects with centroids >154 pixels (remove strong lens objects).
# exclude strong lens objects <=154 distance # The shape of lsst.fits file is 3998,3998 and center is 1699,1699. df['x_center'] = 1699 df['y_center'] = 1699 df['distance'] = ( (df['x[0]'] - df['x_center'])**2 + (df['x[1]'] - df['y_center'])**2 )**0.5 df = df[df.distance > 154]
Imcat script
# create new columns and cleaning (four files)
lc -C -n fN -n id -N '1 2 x' -N '1 2 errx' -N '1 2 g' -n ellip -n flux -n radius < "${M9T}".txt | lc +all 'mag = %flux log10 -2.5 *' | cleancat 15 | lc +all -r 'mag' > "${M9C}".cat
# merge 4 catalogs
mergecats 5 "${MC}".cat "${M9C}".cat "${LC}".cat "${L9C}".cat > ${catalogs}/merge.cat &&
lc -b +all
'x = %x[0][0] %x[1][0] + %x[2][0] + %x[3][0] + 4 / %x[0][1] %x[1][1] + %x[2][1] + %x[3][1] + 4 / 2 vector'
'gm = %g[0][0] %g[1][0] + 2 / %g[0][1] %g[1][1] + 2 / 2 vector'
'gc = %g[2][0] %g[3][0] + 2 / %g[2][1] %g[3][1] + 2 / 2 vector'
'gmd = %g[0][0] %g[1][0] - 2 / %g[0][1] %g[1][1] - 2 / 2 vector'
'gcd = %g[2][0] %g[3][0] - 2 / %g[2][1] %g[3][1] - 2 / 2 vector'
< ${catalogs}/merge.cat > ${final}/final_${i}.cat
Notes
final_text.txt is created by imcat program after merging four lsst files (m,m9,l,l9) after cleaning.
Imports¶
In [1]:
import json, os,sys
import numpy as np
import pandas as pd
import seaborn as sns
import time
sns.set(color_codes=True)
import plotly
import ipywidgets
pd.set_option('display.max_columns',200)
time_start_notebook = time.time()
import matplotlib.pyplot as plt
plt.style.use('ggplot')
%matplotlib inline
print([(x.__name__, x.__version__) for x in [np,pd,sns,plotly,ipywidgets]])
In [2]:
%%javascript
IPython.OutputArea.auto_scroll_threshold = 9999;
Load the final text cleancat15 data¶
g_sq = g00 g00 + g10 g10
gmd_sq = gmd0**2 + gmd1**2In [3]:
!head -2 ../data/cleancat/final_text_cleancat15_100.txt
In [4]:
names = "fN[0][0] fN[1][0] fN[2][0] fN[3][0] id[0][0] id[1][0] id[2][0] id[3][0] x[0] x[1] errx[0][0] errx[0][1] errx[1][0] errx[1][1] errx[2][0] errx[2][1] errx[3][0] errx[3][1] g[0][0] g[0][1] g[1][0] g[1][1] g[2][0] g[2][1] g[3][0] g[3][1] ellip[0][0] ellip[1][0] ellip[2][0] ellip[3][0] flux[0][0] flux[1][0] flux[2][0] flux[3][0] radius[0][0] radius[1][0] radius[2][0] radius[3][0] mag[0][0] mag[1][0] mag[2][0] mag[3][0] gm[0] gm[1] gc[0] gc[1] gmd[0] gmd[1] gcd[0] gcd[1]"
print(names)
In [5]:
names = ['fN[0][0]','fN[1][0]','fN[2][0]','fN[3][0]',
'id[0][0]','id[1][0]','id[2][0]','id[3][0]',
'x[0]','x[1]',
'errx[0][0]','errx[0][1]','errx[1][0]','errx[1][1]','errx[2][0]',
'errx[2][1]','errx[3][0]','errx[3][1]',
'g[0][0]','g[0][1]','g[1][0]','g[1][1]','g[2][0]','g[2][1]','g[3][0]','g[3][1]',
'ellip[0][0]','ellip[1][0]','ellip[2][0]','ellip[3][0]',
'flux[0][0]','flux[1][0]','flux[2][0]','flux[3][0]',
'radius[0][0]','radius[1][0]','radius[2][0]','radius[3][0]',
'mag[0][0]','mag[1][0]','mag[2][0]','mag[3][0]',
'gm[0]','gm[1]','gc[0]', 'gc[1]',
'gmd[0]','gmd[1]','gcd[0]','gcd[1]']
file_path = '../data/cleancat/final_text_cleancat15_000_099.txt'
df = pd.read_csv(file_path,comment='#',engine='python',sep=r'\s\s+',
header=None,names=names)
print(df.shape)
# new columns
# df['g_sq'] = df['g[0][0]'] **2 + df['g[1][0]']**2 # only for imcat 00 and 10
# df['gmd_sq'] = df['gmd[0]'] **2 + df['gmd[1]']**2
df['g_sq'] = df['g[0][0]'] **2 + df['g[0][1]']**2
df['gmd_sq'] = df['gmd[0]'] **2 + df['gmd[1]']**2
df['gm_sq'] = df['gm[0]']**2 + df['gm[1]']**2
df['gc_sq'] = df['gc[0]']**2 + df['gc[1]']**2
df['mag_mono'] = (df['mag[0][0]'] + df['mag[1][0]'] ) / 2
df['mag_chro'] = (df['mag[2][0]'] + df['mag[3][0]'] ) / 2
df.head()
Out[5]:
Plot g-squared for monochromatic and chromatic files¶
In [6]:
fig,ax = plt.subplots(1,2,figsize=(14,6))
df['gm_sq'].plot.hist(bins=50,ax=ax[0],color='b',label='mono')
ax[0].set_xlabel('gm_sq')
ax[0].legend()
ax[0].set_title('g-squared histogram for mono')
# gcsq
df['gc_sq'].plot.hist(bins=50,ax=ax[1],color='r',label='chro')
ax[1].set_xlabel('gc_sq')
ax[1].legend()
ax[1].set_title('g-squared histogram for chro')
# savefig
plt.savefig('images/gmsq_gcsq_hist.png')
In [7]:
plt.figure(figsize=(12,8))
sns.distplot(df['g_sq'],label='g_sq')
sns.distplot(df['gmd_sq'],label='gmd_sq')
plt.legend()
Out[7]:
Contour Plots¶
In [8]:
import plotly
import plotly.graph_objs as go
import plotly.figure_factory as ff
import plotly.tools as tls
from plotly.offline import plot, iplot, init_notebook_mode
init_notebook_mode(connected=False)
In [9]:
def matrix_of_number_density_from_two_cols(df,xcol,ycol,N):
"""Create grid of number density of two columns
- Find the absolute max from two columns.
- Create N bins -absMax to +absMax.
- Return a matrix of N*N shape.
"""
from itertools import product
# derived variables
xlabel = xcol
ylabel = ycol
xlabel1 = xlabel + '_cat_labels'
ylabel1 = ylabel + '_cat_labels'
xlabel2 = xlabel + '_cat'
ylabel2 = ylabel + '_cat'
colname = 'cat_freq'
# take only xlabel and ylabel columns
dx = df[[xlabel, ylabel]].copy()
# get absolute maximum from two columns
tolerance = 0.0000001
max_abs_xcol_ycol = df[[xcol,ycol]].describe().iloc[[3,-1],:].abs().max().max()
# create 1d array with N+1 values to create N bins
# bins = np.linspace(-max_abs_xcol_ycol-tolerance, max_abs_xcol_ycol+tolerance,N+1)
bins = np.linspace(0, max_abs_xcol_ycol,N+1)
# create N bins
dx[xlabel1] = pd.cut(dx[xlabel], bins, labels=np.arange(N))
dx[ylabel1] = pd.cut(dx[ylabel], bins, labels=np.arange(N))
# count number of points in each bin
dx[colname] = dx.groupby([xlabel1,ylabel1])[xlabel1].transform('count')
# drop duplicates
dx1 = dx.drop_duplicates(subset=[xlabel1,ylabel1])[[xlabel1,ylabel1,colname]]
# use permutation to get the grid of N * N
perms = list(product(range(N), range(N)))
x = [i[0] for i in perms]
y = [i[1] for i in perms]
dx2 = pd.DataFrame({xlabel1: x, ylabel1: y, colname:0})
# update dx2 to merge frequency values
dx2.update(dx2.drop(colname,1).merge(dx1,how='left'))
dx2 = dx2.astype(int)
# z values to plot heatmap
z = dx2[colname].values.reshape(N,N)
z = z.T
return z
Transform and scale the data¶
In [10]:
def transform_scale(z,transform='linear',scale=None):
"""Transform and scale given 1d numpy array.
transform: linear, log, sqrt, sinh, arcsinh
scale : minmax, zscale
"""
#==================================
# linear, log, sqrt, arcsinh, sinh
#
# we need linear tranform option to compare.
if transform == 'linear':
z = z
if transform == 'log':
z = np.log1p(z)
if transform == 'sqrt':
z = np.sqrt(z)
if transform == 'sinh':
z = np.sinh(z)
if transform == 'arcsinh':
z = np.arcsinh(z)
#===============================
# scaling minmax and zscale
if scale== 'minmax':
z = z / (z.max()-z.min())
if scale == 'zscale':
z = (z-z.mean()) / z.std()
return z
plot the countours¶
In [11]:
def plot_contour(Z,colorscale,x1y1x2y2=None,
title='Contour plot',
xlabel='',
ylabel=''):
"""Plot the contour plot.
Contour plot from given 2d array.
Example:
-----------
z = matrix_of_number_density_from_two_cols(df,'gsq','gmdsq',N)
z1 = transform_scale(z,transform=transform,scale=scale)
plot_contour(z1, colorscale='Viridis', x1y1x2y2=[10,0,99,90],
title=f'Contour plot: {transform}+{scale}',
xlabel='Bin number of gsq',
ylabel='Bin number of gmsq')
"""
trace1 = go.Contour(z=Z, colorscale=colorscale)
axis_layout = dict(
showticklabels = True
)
xaxis = {**axis_layout, **{'title':f'{xlabel}'}}
yaxis = {**axis_layout, **{'title':f'{ylabel}'}}
layout = go.Layout(
width=1000,
height=1000,
autosize=False,
title=f'{title} ',
xaxis = xaxis,
yaxis = yaxis,
)
data = [trace1]
if x1y1x2y2:
# center line
p2x1, p2y1 = 0,0
p2x2, p2y2 = 99,99
sc1 = go.Scatter(x=[p2x1,p2x2],
y=[p2y1,p2y2],
mode = 'lines+markers',
name = f'line ({p2x1},{p2y1}) to ({p2x2},{p2y2})')
sc2 = go.Scatter(x=[x1y1x2y2[0],x1y1x2y2[2]],
y=[x1y1x2y2[1],x1y1x2y2[3]],
mode = 'lines+markers',
name = f'line ({x1y1x2y2[0]},{x1y1x2y2[1]}) \
to ({x1y1x2y2[2]},{x1y1x2y2[3]})')
data = [trace1, sc1, sc2]
fig = go.Figure( data=data, layout=layout )
# update figure
fig.update_layout(
xaxis = dict(tickmode = 'linear',dtick = 5),
yaxis = dict(tickmode = 'linear',dtick = 5),
)
iplot(fig)
return None
N = 100
transform='log'
scale='zscale'
xcol,ycol = 'g_sq', 'gmd_sq'
z = matrix_of_number_density_from_two_cols(df,xcol,ycol,N)
z1 = transform_scale(z,transform=transform,scale=scale)
plot_contour(z1, colorscale='Viridis', x1y1x2y2=[10,0,99,90],
title=f'Contour plot: {transform}+{scale}',
xlabel=f'Bin number of {xcol}',
ylabel=f'Bin number of {ycol}')
Grid of N*N from g_sq and gmd_sq¶
In [12]:
N = 100
xcol = 'g_sq'
ycol = 'gmd_sq'
max_abs_xcol_ycol = df[[xcol,ycol]].abs().max().max()
print(max_abs_xcol_ycol)
df[df[xcol]==max_abs_xcol_ycol]
Out[12]:
In [13]:
# create 1d array with N+1 values to create N bins
bins = np.linspace(0, max_abs_xcol_ycol,N+1)
# bins dict
bins_dict = {i:v for i,v in enumerate(bins)}
# create N bins
ser_ycol_bins = pd.cut(df[ycol], bins=bins,)
df['g_sq_bins'] = ser_ycol_bins
df[['g_sq','g_sq_bins']].head()
Out[13]:
In [14]:
# bin 0 ==> 0 to 0.03994369
# bin 1 ==> 0.03994369 to 0.07988739
# bin 10 ==> 0.39943694 to 0.43938063
bins[10], bins[11]
Out[14]:
Analysis of Points above gmd_sq bin number 10¶
What is the corresponding gmsq value to y-axis bin number 10 (11th bin)?
The 100*100 bin is created from absMax of g_sq and gmd_sq. Then the line 0 to absMax is divided into 100 parts and bins are created.
bin 10 for gmd_sq is 0.39943694 to 0.43938063
In [15]:
# take all points where gmd_sq > 10th bin
df_gmd_sq_lt_bin10 = df[df.gmd_sq > bins[10]]
print(df_gmd_sq_lt_bin10.shape)
df_gmd_sq_lt_bin10.head()
Out[15]:
In [16]:
# fN[0][0] ==> lsst_mono_z1.5_000.fits
# fN[1][0] ==> lsst_mono90_z1.5_000.fits
#
# id[0][0] ==> id of given object for mono file number 0
In [17]:
# take only first file number 0 (it has m,m9,c and c9)
df_gmd_sq_lt_bin10_file0 = df_gmd_sq_lt_bin10[df_gmd_sq_lt_bin10['fN[0][0]'] == 0]
# add radius for mono
df_gmd_sq_lt_bin10_file0['radius_mono'] = \
(df_gmd_sq_lt_bin10_file0['radius[0][0]'] +
df_gmd_sq_lt_bin10_file0['radius[1][0]'] ) /2.0
print(df_gmd_sq_lt_bin10_file0.shape)
df_gmd_sq_lt_bin10_file0.head()
Out[17]:
Regions above and below for gsq vs gmdsq contour plot¶
For example, take points:
Lower line below the diagonal line
point on x-axis: x1,y1 = 10,0
point on y-axis: x2,y2 = 99,90
here 20,0,99,80 are bin number, their values are obtained from bins_dict
x1,y1 = bins_dict[10], bins_dict[0]
x2,y2 = bins_dict[99], bins_dict[90]
Equation of straight line:
y-y1 = y2-y1 * (x-x1)
-----
x2-x1
boundary: (x2-x1) * (y-y1) - (y2-y1) * (x-x1)In [18]:
N = 100
bins = np.linspace(0, max_abs_xcol_ycol,N+1)
bins_dict = {i:v for i,v in enumerate(bins)}
# points from contour plot
x1y1x2y2 = [10,0,99,90]
print(x1y1x2y2)
# actual values of x1, y1, x2, and y2
x1,y1 = bins_dict[x1y1x2y2[0]], bins_dict[x1y1x2y2[1]]
x2,y2 = bins_dict[x1y1x2y2[2]], bins_dict[x1y1x2y2[3]]
df['above'] = df.eval(" ( (@x2-@x1) * gmd_sq ) >= ( (@y2-@y1) * (g_sq - @x1 )) ")
print(df['above'].value_counts())
print()
print(df['above'].value_counts(normalize=True))
In [19]:
df[df.above==True]['gm_sq'].plot.hist(figsize=(12,8),bins=60,color='b',label='Above')
nabove = len(df[df.above==True])
nbelow = len(df[df.above==False])
above_pct = nabove/(nabove+nbelow)*100
below_pct = 100-above_pct
text = f'Above = {nabove:,} ({above_pct:,.0f}%)\nBelow = {nbelow:,} ( {below_pct:,.0f}%)'
plt.text(0.5,10_000,text,fontsize=14)
plt.legend()
plt.xlabel('gm_sq')
plt.title('gm_sq histograms for above and below cases', fontsize=20)
df[df.above==False]['gm_sq'].plot.hist(figsize=(12,8),bins=60,color='r',alpha=0.5,label='Below')
plt.legend()
plt.xlabel('gm_sq')
plt.title('gm_sq histograms below the boundary', fontsize=20)
Out[19]:
In [20]:
df[df.above==True]['gm_sq'].plot.hist(figsize=(12,8),bins=60,color='r',alpha=0.5)
plt.xlabel('gm_sq')
plt.title('gm_sq histogram', fontsize=20)
Out[20]:
In [21]:
df[df.above==False]['gm_sq'].plot.hist(figsize=(12,8),bins=60,color='r',alpha=0.5,label='Below')
plt.legend()
plt.xlabel('gm_sq')
plt.title('gm_sq histograms below the boundary', fontsize=20)
Out[21]:
Now, Take only the data above the boundary as cleaned data¶
In [22]:
rows_before = df.shape[0]
df = df[df.above==True]
print(f'Before rows: {rows_before}, After rows: {df.shape[0]}')
In [23]:
plt.figure(figsize=(12,8))
sns.distplot(df['gm_sq'],label='gm_sq')
plt.legend()
plt.savefig('images/gmsq_histogram_after_cleaning.png',dpi=300)
gm vs gc Plots¶
Equation of straight line:
y-y1 = y2-y1 * (x-x1)
-----
x2-x1
boundary: (x2-x1) * (y-y1) - (y2-y1) * (x-x1)In [24]:
fig,ax = plt.subplots(1,2,figsize=(16,8))
df.plot.scatter(x='gm[0]',y='gc[0]', ax=ax[0])
df.plot.scatter(x='gm[1]',y='gc[1]', ax=ax[1])
ax[0].set_xlim(-2.0,2.0)
ax[0].set_ylim(-2.0,2.0)
ax[1].set_xlim(-2.0,2.0)
ax[1].set_ylim(-2.0,2.0)
# 45 degree line
n=2.0
ax[0].plot([-n,n],[-n,n],'r--')
ax[1].plot([-n,n],[-n,n],'r--')
plt.suptitle('gm vs gc plot',weight='bold',fontsize=24);
Remove outliers based on gm vs gc plot¶
In [25]:
# Two scatter plots
#===========================================
fig,ax = plt.subplots(1,2,figsize=(16,8))
plt.suptitle('gm vs gc plot',weight='bold',fontsize=24);
df.plot.scatter(x='gm[0]',y='gc[0]', ax=ax[0])
df.plot.scatter(x='gm[1]',y='gc[1]', ax=ax[1])
ax[0].set_xlim(-2.0,2.0)
ax[0].set_ylim(-2.0,2.0)
ax[1].set_xlim(-2.0,2.0)
ax[1].set_ylim(-2.0,2.0)
# 45 degree line
n=2.0
ax[0].plot([-n,n],[-n,n],'r--')
ax[1].plot([-n,n],[-n,n],'r--')
# Left plot gm0 vs gc0 parallel lines
#===========================================
x = df['gm[0]'].to_numpy()
y = df['gc[0]'].to_numpy()
# Find distance for parallel lines
sigma = (x - y).std()
n = 10
d = n * sigma
c = d/np.sqrt(2)
y1 = x + c
y2 = x - c
# left plot parallel lines
ax[0].plot(y1,x,'b-.',label=f'{n} sigma')
ax[0].plot(y2,x,'b-.')
# Right plot gm1 vs gc1 parallel lines
#===========================================
x = df['gm[1]'].to_numpy()
y = df['gc[1]'].to_numpy()
sigma = (x - y).std()
n = 10
d = n * sigma
c = d/np.sqrt(2)
y1 = x + c
y2 = x - c
ax[1].plot(y1,x,'b-.', label=f'{n} sigma')
ax[1].plot(y2,x,'b-.')
#=============
# Add legends
# plt.tight_layout()
ax[0].legend(loc='upper left')
ax[1].legend(loc='upper left')
plt.show()
In [26]:
# Remove outliers for gc0 vs gm0
n = 10
sigma = (df['gm[0]'] - df['gc[0]']).std()
d = n * sigma
c = d/np.sqrt(2)
cond_upper = (df['gc[0]'] - df['gm[0]'] <= c)
cond_below = (df['gm[0]'] - df['gc[0]'] <= c)
df = df[ cond_upper & cond_below ]
# df.plot.scatter(x='gm[0]',y='gc[0]')
In [27]:
# Remove outliers for gc1 vs gm1
n = 10
sigma = (df['gm[1]'] - df['gc[1]']).std()
d = n * sigma
c = d/np.sqrt(2)
cond_upper = (df['gc[1]'] - df['gm[1]'] <= c)
cond_below = (df['gm[1]'] - df['gc[1]'] <= c)
df = df[ cond_upper & cond_below ]
# df.plot.scatter(x='gm[1]',y='gc[1]')
In [28]:
# Two scatter plots
#===========================================
fig,ax = plt.subplots(1,2,figsize=(16,8))
plt.suptitle('gm vs gc plot',weight='bold',fontsize=24);
df.plot.scatter(x='gm[0]',y='gc[0]', ax=ax[0])
df.plot.scatter(x='gm[1]',y='gc[1]', ax=ax[1])
ax[0].set_xlim(-2.0,2.0)
ax[0].set_ylim(-2.0,2.0)
ax[1].set_xlim(-2.0,2.0)
ax[1].set_ylim(-2.0,2.0)
# 45 degree line
n=2.0
ax[0].plot([-n,n],[-n,n],'r--')
ax[1].plot([-n,n],[-n,n],'r--')
# Left plot gm0 vs gc0 parallel lines
#===========================================
x = df['gm[0]'].to_numpy()
y = df['gc[0]'].to_numpy()
# Find distance for parallel lines
sigma = (x - y).std()
n = 10
d = n * sigma
c = d/np.sqrt(2)
y1 = x + c
y2 = x - c
# left plot parallel lines
ax[0].plot(y1,x,'b-.',label=f'{n} sigma')
ax[0].plot(y2,x,'b-.')
# Right plot gm1 vs gc1 parallel lines
#===========================================
x = df['gm[1]'].to_numpy()
y = df['gc[1]'].to_numpy()
sigma = (x - y).std()
n = 10
d = n * sigma
c = d/np.sqrt(2)
y1 = x + c
y2 = x - c
ax[1].plot(y1,x,'b-.', label=f'{n} sigma')
ax[1].plot(y2,x,'b-.')
#=============
# Add legends
# plt.tight_layout()
ax[0].legend(loc='upper left')
ax[1].legend(loc='upper left')
plt.show()
Plot magnidude for different bin numbers¶
In [29]:
df.filter(regex='mag').head(2)
Out[29]:
In [30]:
def plot_bin_mag_mono_chro(nbins,show=False):
import os
if not os.path.isdir('images'):
os.makedirs('images')
df['bins_mag_mono'] = pd.cut(df['mag_mono'],nbins)
df['bins_mag_chro'] = pd.cut(df['mag_chro'],nbins)
text_mono = df.groupby('bins_mag_mono')['gm_sq'].count().to_string()
text_chro = df.groupby('bins_mag_chro')['gc_sq'].count().to_string()
# plot
fig,ax = plt.subplots(1,2,figsize=(12,8))
# mono
df.groupby('bins_mag_mono')['gm_sq'].mean().plot(marker='o',ax=ax[0])
ax[0].tick_params(axis='x', rotation=90)
ax[0].set_ylabel('gm_sq_mean',fontsize=18)
ax[0].set_xlabel('bin_mag_mono',fontsize=18)
ax[0].set_title(f'gm_sq per magnitude bins with nbins = {nbins}')
ax[0].text(0,0.5,text_mono,fontsize=14,va='center')
ax[0].set_ylim(0,1)
ax[0].set_yticks(np.arange(0, 1, step=0.1))
# chro
df.groupby('bins_mag_chro')['gc_sq'].mean().plot(marker='o',ax=ax[1])
ax[1].tick_params(axis='x', rotation=90)
ax[1].set_ylabel('gc_sq_mean',fontsize=18)
ax[1].set_xlabel('bin_mag_chro',fontsize=18)
ax[1].set_title(f'gc_sq per magnitude bins with nbins = {nbins}')
ax[1].text(0,0.5,text_chro,fontsize=14,va='center')
ax[1].set_ylim(0,1)
ax[1].set_yticks(np.arange(0, 1, step=0.1))
plt.tight_layout()
plt.savefig(f'images/bin_mag_mono_chro_{nbins}.png')
if show:
plt.show()
plt.close()
for nbins in range(5,15):
plot_bin_mag_mono_chro(nbins,show=True)
Magnitude weight column for Monochromatic case¶
In [31]:
df['mag_mono'].plot.hist()
Out[31]:
In [32]:
from scipy.optimize import curve_fit
# look at case when nbins = 9 and when the curve is going up
mag_low_nbins9 = (-12.884-12.028) / 2
mag_high_nbins9 = (-10.314-9.457) / 2
xcol = 'mag_mono'
ycol = 'gm_sq'
x = df.query(""" @mag_low_nbins9 < mag_mono < @mag_high_nbins9 """)[xcol].to_numpy()
y = df.query(""" @mag_low_nbins9 < mag_mono < @mag_high_nbins9 """)[ycol].to_numpy()
def func(x, a, b):
return a*x + b
params, _ = curve_fit(func, x, y)
[a, b] = params.round(2)
print(f'magnitude ranges for mono: {mag_low_nbins9}, {mag_high_nbins9}')
print(f'fitting params for mono: {a}, {b}' )
In [33]:
def magnitude_weight_mono(mag):
if mag < -12.456:
return 1/ (-12.456 *a + b)
else:
return 1/ (a*mag + b)
df['wt_mag_mono'] = df['mag_mono'].apply(magnitude_weight_mono)
df['wt_mag_mono'] = df['wt_mag_mono'] / df['wt_mag_mono'].mean() # normalize by mean
Magnitude weight column for Chromatic case¶
In [34]:
df['mag_chro'].plot.hist()
Out[34]:
In [35]:
from scipy.optimize import curve_fit
# look at case when nbins = 9 and when the curve is going up
mag_low_nbins9 = (-12.895-12.041) / 2
mag_high_nbins9 = (-10.333-9.479) / 2
xcol = 'mag_chro'
ycol = 'gc_sq'
x = df.query(""" @mag_low_nbins9 < mag_chro < @mag_high_nbins9 """)[xcol].to_numpy()
y = df.query(""" @mag_low_nbins9 < mag_chro < @mag_high_nbins9 """)[ycol].to_numpy()
def func(x, a, b):
return a*x + b
params, _ = curve_fit(func, x, y)
[a, b] = params.round(2)
print(f'magnitude ranges for chro: {mag_low_nbins9}, {mag_high_nbins9}')
print(f'fitting params for chro: {a}, {b}' )
In [36]:
def magnitude_weight_chro(mag):
if mag < -12.468:
return 1/ (-12.468*a + b)
else:
return 1/ (a*mag + b)
df['wt_mag_chro'] = df['mag_chro'].apply(magnitude_weight_chro)
df['wt_mag_chro'] = df['wt_mag_chro'] / df['wt_mag_chro'].mean() # normalize by mean
# mean
df['wt_mag'] = (df['wt_mag_mono'] + df['wt_mag_chro']) / 2
# df.drop(['wt_mag_chro','wt_mag_mono'],axis=1,inplace=True)
df.iloc[:2,-7:]
Out[36]:
Ellipticity Components Transformation¶
c2 = (dx * dx - dy * dy) / (r * r);
s2 = 2 * dx * dy / (r * r);
eX = s2 * e[0] + c2 * e[1];
eesum += eX * eX * w[0] * w[0];
eTsum[bin] -= (c2 * e[0] + s2 * e[1]) * w[0];
In [37]:
df.head(2)
Out[37]:
In [38]:
# constants
RMIN = 10
DLNR = 0.5
df['dx'] = df['x[0]'] - 1699 # jesisim fitsfiles have shape 3398, 3398
df['dy'] = df['x[1]'] - 1699
df['r'] = np.hypot(df['dx'], df['dy'])
# df['r'] = np.sqrt(df['dx']**2 + df['dy']**2)
df['cos2theta'] = df.eval(' (dx * dx - dy * dy) / (r * r)' )
df['sin2theta'] = df.eval(' (2 * dx * dy ) / (r * r)' )
df['bin'] = ( np.log(df.r / RMIN) / DLNR).astype(int)
df['bin'].value_counts()
Out[38]:
In [39]:
df['eX_mono'] = df['sin2theta'] * df['gm[0]'] + df['cos2theta'] * df['gm[1]']
df['eT_mono'] = -1 * (df['cos2theta'] * df['gm[0]'] + df['sin2theta'] * df['gm[1]'] )
df['eX_chro'] = df['sin2theta'] * df['gc[0]'] + df['cos2theta'] * df['gc[1]']
df['eT_chro'] = -1 * (df['cos2theta'] * df['gc[0]'] + df['sin2theta'] * df['gc[1]'] )
In [40]:
df['eT_mono_times_wt'] = df['eT_mono'] * df['wt_mag']
df['eT_chro_times_wt'] = df['eT_chro'] * df['wt_mag']
In [41]:
df.head()
Out[41]:
Radial Bins groupings¶
In [42]:
df_radial_bins = df.groupby('bin').agg({'r': 'mean',
'wt_mag': 'sum',
'eT_mono_times_wt': 'sum',
'eT_chro_times_wt': 'sum',
})
df_radial_bins.columns = ['r_mean',
'wt_mag_sum',
'eT_mono_times_wt_sum',
'eT_chro_times_wt_sum']
df_radial_bins['eT_mean_mono'] = df_radial_bins.eval('eT_mono_times_wt_sum / wt_mag_sum')
df_radial_bins['eT_mean_chro'] = df_radial_bins.eval('eT_chro_times_wt_sum / wt_mag_sum')
df_radial_bins['bin_count'] = df['bin'].value_counts()
print('Statistics for different radial bins')
print(f'RMIN = {RMIN} and DLNR = {DLNR}')
df_radial_bins.style\
.background_gradient(subset=['eT_mean_mono','eT_mean_chro'],cmap='Blues')\
.apply(lambda x: ["background-color: #DAabaa"
if (v < 0)
else "" for i, v in enumerate(x)], axis = 1)
Out[42]:
In [43]:
# why some eT values are -ve?
"""
1. For given rmin and dlnr we have some bins very few object counts.
""";
In [44]:
pd.cut(df['eT_mono'],20).value_counts()
Out[44]:
In [45]:
fig,ax = plt.subplots(figsize=(12,8))
sns.distplot(df['eT_mono'],ax=ax)
plt.title('Histogram and density plot of eT_mono')
Out[45]:
In [46]:
df['wt_mag'].hist(bins=100,figsize=(12,8))
plt.xlim(0,3)
Out[46]:
In [47]:
pd.options.display.max_columns=500
df.head(5)
Out[47]:
In [48]:
df.query(""" r>500 """)['eT_mono'].describe()
Out[48]:
In [49]:
df['gm_sq'].hist(bins=100,figsize=(12,8))
Out[49]:
Plot of eT for mono and chro¶
In [50]:
def plot_eT_mean(rmin, dlnr,min_bin_count,title='',
show_mono_chro=False,
show_data=True):
df['dx'] = df['x[0]'] - 1699 # jesisim fitsfiles have shape 3398, 3398
df['dy'] = df['x[1]'] - 1699
df['r'] = np.hypot(df['dx'], df['dy'])
df['cos2theta'] = df.eval(' (dx * dx - dy * dy) / (r * r)' )
df['sin2theta'] = df.eval(' (2 * dx * dy ) / (r * r)' )
df['bin'] = ( np.log(df.r / rmin) / dlnr).astype(int)
df_radial_bins = df.groupby('bin').agg({'r': 'mean',
'wt_mag': 'sum',
'eT_mono_times_wt': 'sum',
'eT_chro_times_wt': 'sum',
})
df_radial_bins.columns = ['r_mean',
'wt_mag_sum',
'eT_mono_times_wt_sum',
'eT_chro_times_wt_sum']
df_radial_bins['eT_mean_mono'] = df_radial_bins.eval('eT_mono_times_wt_sum / wt_mag_sum')
df_radial_bins['eT_mean_chro'] = df_radial_bins.eval('eT_chro_times_wt_sum / wt_mag_sum')
# eT mono vs chro ratio
df_radial_bins['eT_ratio'] = df_radial_bins['eT_mean_mono'] / df_radial_bins['eT_mean_chro']
# bin counts
df_radial_bins['bin_count'] = df['bin'].value_counts()
# init figure
fig, ax = plt.subplots(1,1,figsize=(12,8))
# min bin count
df_radial_bins = df_radial_bins.query(""" bin_count > @min_bin_count """)
# plot ratio of eT mono vs chro
df_radial_bins.plot.line(x='r_mean',y='eT_ratio',marker='o',color='b',ax=ax)
# also plot eT mono and eT chro
if show_mono_chro:
df_radial_bins.plot.line(x='r_mean',y='eT_mean_mono',marker='o',color='r', ax=ax)
df_radial_bins.plot.line(x='r_mean',y='eT_mean_chro',marker='o',color='b',ax=ax)
# label
plt.ylabel('eT')
plt.title(title)
plt.suptitle('Plot of eT',fontsize=24,weight='bold')
# also show the dataframe
if show_data:
display(df_radial_bins.style.background_gradient(subset=['eT_mean_mono']))
plt.show()
rmin = 300
dlnr = 0.5
min_bin_count = 10
title = f'rmin={rmin}, dlnr={dlnr}, min_bin_count={min_bin_count}'
plot_eT_mean(rmin=rmin, dlnr=dlnr,min_bin_count=min_bin_count,title=title)
Interactive plots¶
In [51]:
import plotly.graph_objs as go
import plotly.figure_factory as ff
from plotly import tools
from plotly.offline import plot, iplot, init_notebook_mode
init_notebook_mode(connected=False)
In [52]:
rmin = 400
dlnr = 0.4
min_bin = 20
siingle = False
In [53]:
def plotly_eT_plot(rmin=400, dlnr=0.4, min_bin=20,
show_mono_chro=False,
show_data=True,
title=f'rmin={rmin}, dlnr={dlnr}, min_bin={min_bin}'):
df['dx'] = df['x[0]'] - 1699 # jesisim fitsfiles have shape 3398, 3398
df['dy'] = df['x[1]'] - 1699
df['r'] = np.hypot(df['dx'], df['dy'])
df['cos2theta'] = df.eval(' (dx * dx - dy * dy) / (r * r)' )
df['sin2theta'] = df.eval(' (2 * dx * dy ) / (r * r)' )
df['bin'] = ( np.log(df.r / rmin) / dlnr).astype(int)
df_radial_bins = df.groupby('bin').agg({'r': 'mean',
'wt_mag': 'sum',
'eT_mono_times_wt': 'sum',
'eT_chro_times_wt': 'sum',
})
df_radial_bins.columns = ['r_mean',
'wt_mag_sum',
'eT_mono_times_wt_sum',
'eT_chro_times_wt_sum']
df_radial_bins['eT_mean_mono'] = df_radial_bins.eval('eT_mono_times_wt_sum / wt_mag_sum')
df_radial_bins['eT_mean_chro'] = df_radial_bins.eval('eT_chro_times_wt_sum / wt_mag_sum')
# ratio of eT mono and chro
df_radial_bins['eT_ratio'] = df_radial_bins['eT_mean_mono'] / df_radial_bins['eT_mean_chro']
df_radial_bins['bin_count'] = df['bin'].value_counts()
# min bin count
df_radial_bins = df_radial_bins.query(""" bin_count > @min_bin """)
# eT mono vs chro ratio
sc = go.Scatter( x=df_radial_bins['r_mean'],
y=df_radial_bins['eT_ratio'],
mode = 'lines+markers',
name = 'eT ratio')
mydata = [sc]
# also plot eT mono and chro
if show_mono_chro:
sc1 = go.Scatter(x=df_radial_bins['r_mean'],
y=df_radial_bins['eT_mean_mono'],
mode = 'lines+markers',
name = 'eT mono')
sc2 = go.Scatter(x=df_radial_bins['r_mean'],
y=df_radial_bins['eT_mean_chro'],
mode = 'lines+markers',
name = 'eT chro')
mydata= [sc1,sc2]
# layout
layout = go.Layout(
title=title,
xaxis=dict(
title='radius',
titlefont=dict(
family='Courier New, monospace',
size=18,
color='#7f7f7f')),
yaxis=dict(title='eT',
titlefont=dict(
family='Courier New, monospace',
size=18,
color='#7f7f7f')))
myfig = go.Figure(data=mydata, layout=layout)
myfig.update_layout(showlegend=True)
# also show the dataframe
if show_data:
display(df_radial_bins.style.background_gradient(subset=['eT_mean_mono']))
# iplot plots in jupyter-notebook, plot opens in new tab.
return iplot(myfig, filename='myplot.html')
# plot
rmin = 300
dlnr = 0.5
min_bin = 20
title = f'rmin={rmin}, dlnr={dlnr}, min_bin={min_bin}'
plotly_eT_plot(rmin=rmin, dlnr=dlnr,min_bin=min_bin,title=title)
In [54]:
# plot
rmin = 400
dlnr = 0.5
min_bin = 20
title = f'rmin={rmin}, dlnr={dlnr}, min_bin={min_bin}'
plotly_eT_plot(rmin=rmin, dlnr=dlnr,min_bin=min_bin,title=title,show_mono_chro=True)
ipywidgets¶
which pip
pip install ipywidgets
jupyter nbextension enable --py --sys-prefix widgetsnbextensionIn [55]:
import ipywidgets as widgets
from ipywidgets import HBox, VBox
from ipywidgets import interactive
In [56]:
rmin_ = widgets.IntSlider(min=0, max=500, step=50, value=100)
dlnr_ = widgets.FloatSlider(min=0.1,max=0.8,step=0.05,value=0.5)
min_bin_count_ = widgets.IntSlider(min=10, max=600, step=20, value=50)
interactive_plot = interactive(plotly_eT_plot,
rmin=rmin_,
dlnr=dlnr_,
min_bin_count=min_bin_count_)
output = interactive_plot.children[-1]
interactive_plot
Time Taken¶
In [57]:
time_taken = time.time() - time_start_notebook
h,m = divmod(time_taken,60*60)
print('Time taken to run whole notebook: {:.0f} hr '\
'{:.0f} min {:.0f} secs'.format(h, *divmod(m,60)))
In [58]:
import subprocess
subprocess.call(['python', '-m', 'nbconvert', '*.ipynb'])
Out[58]: